Oilpan: rename EAGERLY_SWEEP() as EAGERLY_FINALIZED().

Oilpan: rename EAGERLY_SWEEP() as EAGERLY_FINALIZED().

Renamed to make its purpose clearer; classes are to be annotated as
eagerly finalized when they either need to promptly release ownership of
non-heap objects, or when their destructors have an unavoidable need to
access other heap objects. Both kinds are considered very rare, but more
accurate control of finalization timing is sometimes needed & vital.

Also introduce a transition time version of the macro,
EAGERLY_FINALIZED_WILL_BE_REMOVED() for the places where eager finalization
is needed non-Oilpan once lazy sweeping is enabled. That class annotation
will be removed once the class is always on the heap.

R=haraken
BUG=420515
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=195887

[sof, 2015-05-25 20:57:56]

sigbjornf@opera.com changed reviewers:
+ oilpan-reviews@chromium.org

[sof, 2015-05-25 20:57:56]

please take a look.

[haraken, 2015-05-25 23:24:16]

LGTM

[sof, 2015-05-26 05:15:15]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2015-05-26 05:15:28]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1150863003/1

[commit-bot: I haz the power, 2015-05-26 05:18:42]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=195887

[sof, 2015-05-26 06:35:23]

I went back & forth on whether to use EAGERLY_FINALIZE() or EAGERLY_FINALIZED()
as a class-level declaration. I now notice that the description wasn't updated
to use what I landed on, EAGERLY_FINALIZE();

[haraken, 2015-05-26 13:06:34]

haraken@chromium.org changed reviewers:
+ haraken@chromium.org

[haraken, 2015-05-26 13:06:35]

https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h
File Source/platform/heap/Heap.h (right):

https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h...
Source/platform/heap/Heap.h:1145: class HeapIndexTrait<T, typename
WTF::EnableIf<IsEagerlyFinalizedType<T>::value>::Type> {

Just to confirm:

class A : GarbageCollected<A> {
  EAGERLY_FINALIZE();
  static A* create() {
    return new A; // This will be created in the eagerly sweep heap.
  }
};

class B : class A {
  static B* create() {
    return new B;  // Will this be created in the eagerly sweep heap?
  }
};

My question is whether the SFINAE returns true for sub-classes.

[sof, 2015-05-26 13:08:42]

On 2015/05/26 13:06:35, haraken wrote:
> https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h
> File Source/platform/heap/Heap.h (right):
> 
>
https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h...
> Source/platform/heap/Heap.h:1145: class HeapIndexTrait<T, typename
> WTF::EnableIf<IsEagerlyFinalizedType<T>::value>::Type> {
> 
> Just to confirm:
> 
> class A : GarbageCollected<A> {
>   EAGERLY_FINALIZE();
>   static A* create() {
>     return new A; // This will be created in the eagerly sweep heap.
>   }
> };
> 
> class B : class A {
>   static B* create() {
>     return new B;  // Will this be created in the eagerly sweep heap?
>   }
> };
> 
> My question is whether the SFINAE returns true for sub-classes.

Doesn't the HeapTest unit test answer your question?

[haraken, 2015-05-26 13:13:19]

On 2015/05/26 13:08:42, sof wrote:
> On 2015/05/26 13:06:35, haraken wrote:
> >
https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h
> > File Source/platform/heap/Heap.h (right):
> > 
> >
>
https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h...
> > Source/platform/heap/Heap.h:1145: class HeapIndexTrait<T, typename
> > WTF::EnableIf<IsEagerlyFinalizedType<T>::value>::Type> {
> > 
> > Just to confirm:
> > 
> > class A : GarbageCollected<A> {
> >   EAGERLY_FINALIZE();
> >   static A* create() {
> >     return new A; // This will be created in the eagerly sweep heap.
> >   }
> > };
> > 
> > class B : class A {
> >   static B* create() {
> >     return new B;  // Will this be created in the eagerly sweep heap?
> >   }
> > };
> > 
> > My question is whether the SFINAE returns true for sub-classes.
> 
> Doesn't the HeapTest unit test answer your question?

Yes, sorry about the noise :)

[sof, 2015-05-26 13:19:26]

On 2015/05/26 13:13:19, haraken wrote:
> On 2015/05/26 13:08:42, sof wrote:
> > On 2015/05/26 13:06:35, haraken wrote:
> > >
> https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h
> > > File Source/platform/heap/Heap.h (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/1150863003/diff/1/Source/platform/heap/Heap.h...
> > > Source/platform/heap/Heap.h:1145: class HeapIndexTrait<T, typename
> > > WTF::EnableIf<IsEagerlyFinalizedType<T>::value>::Type> {
> > > 
> > > Just to confirm:
> > > 
> > > class A : GarbageCollected<A> {
> > >   EAGERLY_FINALIZE();
> > >   static A* create() {
> > >     return new A; // This will be created in the eagerly sweep heap.
> > >   }
> > > };
> > > 
> > > class B : class A {
> > >   static B* create() {
> > >     return new B;  // Will this be created in the eagerly sweep heap?
> > >   }
> > > };
> > > 
> > > My question is whether the SFINAE returns true for sub-classes.
> > 
> > Doesn't the HeapTest unit test answer your question?
> 
> Yes, sorry about the noise :)

There are some complexities when faced with Node & EventTarget though atm,
doesn't work for those.
https://codereview.chromium.org/1157933002/ tries to address, but I'm not
entirely happy with it. (== I haven't published the CL yet :) )